The UK is facing fresh concerns over access to advanced artificial intelligence after Anthropic withheld its latest AI model from Britain’s AI Security Institute (AISI) for pre-release safety testing.

Anthropic’s new Claude Mythos 5.1 was initially made available to selected organisations in the United States, while the UK’s government-backed AI safety body was not given equivalent early access. The model is reported to have capabilities aimed at areas including cybersecurity and life sciences.

The decision has raised concerns in Whitehall about whether the UK could lose access to some of the world’s most advanced AI systems and whether geopolitical tensions could increasingly influence the technology sector.

Anthropic has said it is working with the US government on expanding access, while UK officials are reviewing the potential implications for national security, AI testing and Britain’s position in the global artificial intelligence race.

The development highlights a growing challenge for the UK: maintaining close cooperation with major AI companies while ensuring British authorities can independently assess the risks posed by increasingly powerful models.